基于测试数据时间序列的导弹故障预测  

Missile Fault Prediction Method Based on Time Series of Test Data

作　　者：叶玮[1] 郑守铎[1] 温瑞珩[1] 

机构地区：[1]中国人民解放军92941部队

出　　处：《现代防御技术》2014年第3期154-158,共5页Modern Defence Technology

摘　　要：提出基于导弹测试数据的时间序列预测方法。以实际测试数据为例,建立了导弹故障预测模型。在对原数据序列分析基础上进行了趋势拟合,并进行了残差序列的识别及模型拟合,应用经过模型检验和参数检验后得到的组合模型进行参数预测,证明了不论是点预测还是区间预测,都能有效地提高导弹故障预测准确性。A missile fault prediction method based on time series of test data is put forward. Taking actual test data as example, a prediction model of missile fault is developed. Trend fitting is carried out based on the analysis of original data sequence. In addition, the residual error sequence recognition and the model fitting are carried out. The combined model based on model inspection and parameter inspection is used to make parameter estimation, which proves that point prediction and interval prediction can effectively improve missile fault prediction accuracy.

关 键 词：测试数据 时间序列 平稳性 相关系数 

分 类 号：E927[军事—军事装备学] TB114[兵器科学与技术—武器系统与运用工程]